mlsanityData to 7 Oct 2026

Humanoid and legged locomotion world models

26 papers0 this weekHigh confidence in this grouping: the papers form one clear topicHigh confidence in this grouping: the papers form one clear topic · About

Keywords: humanoid, whole-body, locomotion, loco-manipulation, policy

Key facts. Humanoid and legged locomotion world models: 26 papers on arXiv since 2021 (data through 7 Oct 2026). 8 of its 26 mapped papers (31%) cite “RMA: Rapid Motor Adaptation for Legged Robots” (2021, arXiv 2107.04034). 6 of its 26 mapped papers (23%) cite “World Model-based Perception for Visual Legged Locomotion” (2024, arXiv 2409.16784). 1 new paper 17 Sep–7 Oct.

What’s new

17 Sep – 7 Oct 2026

1 new since 17 Sep

  1. LocoWM: High-Precision Locomotion through World-Model-Guided Residual Adaptation

Foundations

Most cited by papers here

  1. RMA: Rapid Motor Adaptation for Legged Robots2021cited by 8 here
  2. World Model-based Perception for Visual Legged Locomotion2024cited by 6 here
  3. Advancing Humanoid Locomotion: Mastering Challenging Terrains with Denoising World Model Learning2024cited by 5 here
  4. Learning Humanoid Locomotion with World Model Reconstruction2025cited by 4 here
  5. HAIC: Humanoid Agile Object Interaction Control via Dynamics-Aware World Model2026cited by 3 here

Lineage

The papers this direction builds on, oldest first.

Each paper on the line cites the one before.

  1. 2018World Modelscited by 5 here
  2. 2018Learning Latent Dynamics for Planning from Pixelscited by 5 here
  3. 2019Dream to Control: Learning Behaviors by Latent Imaginationcited by 8 here
  4. 2020Mastering Atari with Discrete World Modelscited by 7 here
  5. 2023Mastering Diverse Domains through World Modelscited by 6 here
  6. 2023TD-MPC2: Scalable, Robust World Models for Continuous Controlcited by 5 here
  7. 2025Robotic World Model: A Neural Network Simulator for Robust Policy Optimization in Roboticscited by 7 here
  8. 2026HAIC: Humanoid Agile Object Interaction Control via Dynamics-Aware World Modelcited by 3 here
16 more it builds on, not in this chain
  1. 2021RMA: Rapid Motor Adaptation for Legged Robotscited by 8 here
  2. 2021Learning to Walk in Minutes Using Massively Parallel Deep Reinforcement Learningcited by 6 here
  3. 2022Temporal Difference Learning for Model Predictive Controlcited by 6 here
  4. 2022DayDreamer: World Models for Physical Robot Learningcited by 8 here
  5. 2022Scalable Diffusion Models with Transformerscited by 3 here
  6. 2023Diffusion Policy: Visuomotor Policy Learning via Action Diffusioncited by 4 here
  7. 2023TWIST: Teacher-Student World Model Distillation for Efficient Sim-to-Real Transfercited by 4 here
  8. 2024Genie: Generative Interactive Environmentscited by 3 here
  9. 2024Advancing Humanoid Locomotion: Mastering Challenging Terrains with Denoising World Model Learningcited by 5 here
  10. 2024World Model-based Perception for Visual Legged Locomotioncited by 6 here
  11. 2025Learning Humanoid Locomotion with World Model Reconstructioncited by 4 here
  12. 2025Wan: Open and Advanced Large-Scale Video Generative Modelscited by 3 here
  13. 2025V-JEPA 2: Self-Supervised Video Models Enable Understanding, Prediction and Planningcited by 3 here
  14. 2025Motus: A Unified Latent Action World Modelcited by 3 here
  15. 2026Causal World Modeling for Robot Controlcited by 4 here
  16. 2026World Action Models are Zero-shot Policiescited by 3 here

The 24 most cited of 75.

Browse

21 of 26, ranked 6–26

  1. 6MotionWAM: Towards Foundation World Action Models for Real-Time Humanoid Loco-Manipulation20263
  2. 7Ego-Vision World Model for Humanoid Contact Planning20252
  3. 8Hierarchical World Models as Visual Whole-Body Humanoid Controllers20242
  4. 9ω-0: A Latent Predictive World Action Model for Concurrent Humanoid Loco-Manipulation20261
  5. 10DreamPolicy: A Unified World-model Policy for Scalable Humanoid Locomotion20251
  6. 11MetaWorld: Skill Transfer and Composition in a Hierarchical World Model for Grounding High-Level Instructions20260
  7. 12MetaWorld-X: Hierarchical World Modeling via VLM-Orchestrated Experts for Humanoid Loco-Manipulation20260
  8. 13Learning Task-Invariant Properties via Dreamer: Enabling Efficient Policy Transfer for Quadruped Robots20260
  9. 14UniT: Toward a Unified Physical Language for Human-to-Humanoid Policy Learning and World Modeling20260
  10. 15SplitAdapter: Load-Aware Humanoid Loco-Manipulation via Factorized Adaptation20260
All 21
  1. 16SWAP: Symmetric Equivariant World-Model for Agile Robot Parkour20260
  2. 17DynaWM: Dynamics-Aware Distillation with World Model and Momentum Targets for Smooth Locomotion over Continuous Stairs20260
  3. 18LUCID: Latent-Skill Unified Control via Imagined Dynamics for Long-Horizon Humanoid Loco-Manipulation20260
  4. 19GigaBrain-WBC-0.5: A Behavior World Model for Robust Humanoid Whole-Body Tracking with Environment Interaction20260
  5. 20DreamMimic: Learning Visuomotor Whole-Body Loco-Manipulation via World Model20260
  6. 21World-Model-Augmented Visual Locomotion for Humanoids on Foothold-Constrained Terrain20260
  7. 22CAP: Continuously Adaptive Perception-Blind Humanoid Locomotion via Learned Denoising20260
  8. 23DWMP: Leveraging Dual World Models for Humanoid Obstacle Traversal20260
  9. 24WholeBodyWAM: Generalizing Pre-trained World-Action Priors to Humanoid Loco-Manipulation via WBC-Grounded Coordination20260
  10. 25WholeBodyWAM: Learning Whole-Body World Action Models with Scalable Motion Priors20260
  11. 26Humanoid World Models: Open World Foundation Models for Humanoid Robotics20250